Use topology/links and anynetlinks filters with eq/in operators so
store WAN follow-ups return real peer tunnels instead of unscoped dumps.
Co-authored-by: Cursor <cursoragent@cursor.com>
Adds three new SD-WAN checks (wanAppRtpMos/Loss/Jitter) that measure
REAL voice-traffic quality on actual RTP frames via Prisma DPI, not
synthetic link probes. Graded against the WORST 5-minute window so
transient degradation the 24h link-probe averages smooth away
actually surfaces.
Voice-app selection is tenant-configurable via PRISMA_APP_ID_VOICE +
PRISMA_APP_NAME_VOICE (Webex_Calling_RTP recommended for Webex
Calling shops — the Webex-specific DPI signature excludes non-Webex
UDP noise). Legacy PRISMA_APP_ID_RTP_BASE still honored with a
one-time deprecation warning.
Widens the default WAN look-back from 24h to 7 days: per-app metrics
only get datapoints when calls actually happen, so sporadic Webex
Calling stores (3-4 calls/day) need a wider window for worst-window
statistics to be meaningful. Interval picker snaps 7d to 1hour
buckets (168 pts) to keep payloads bounded while preserving
worst-hour granularity. Hard-capped at 7d — beyond that Prisma
downsamples to 1-day buckets and the signal collapses.
Also:
- Client-side concurrency limiter (PRISMA_MAX_INFLIGHT, default 3)
to prevent 429 cascades when /voicediag fans out 10+ parallel
metric fetches
- "View in Prisma UI" deep links in both /phonestatus WAN follow-up
and /voicediag details, threading through a new
integrations/paloalto/urls.js builder
- humanizeMetricUnit maps raw API unit strings ("percentage",
"milliseconds") to display symbols ("%", "ms") to fix
"11.83percentage" leaking to the UI
- getAppAudio envelope distinguishes not-configured / fetch-failed /
no-traffic states so misleading "set env var" messages don't fire
when the real problem is a 429
Co-authored-by: Cursor <cursoragent@cursor.com>
The wanAlarms check was pasting raw Prisma `info` JSON blobs (nested
`vpn_reasons` arrays with element/site/vpnlink ids) into the chat
message field. On a store with 20 NETWORK_ANYNETLINK_DOWN flaps this
produced a wall of unreadable stringified JSON where the actual
signal ("SD-WAN overlay tunnels are flapping") was lost.
Introduces a shared alarmSemantics module that:
- Buckets each code into overlay / physical / device / other so
the check + renderer stay consistent
- Humanizes codes (NETWORK_ANYNETLINK_DOWN → "SD-WAN overlay
tunnel down") with a Title-Cased fallback for unknown codes
- Rolls up (code + severity) tuples so 20 identical alarms show as
a single line with ×20 and a "just now / Nm / Nh / Nd" age
Rewrites wanAlarms.run() to use those helpers + cross-reference the
site's physical link state so operators aren't left wondering why 20
alarms fired while every metric shows green: overlay flaps get a
"physical WAN paths are all up per Link State" clarifier, and
physical alarms point back at the Link State check. The label loses
its hardcoded "(last 1h)" suffix since the alarm window is now
dynamic (defaults to 24h to match the WAN window).
The follow-up renderer used by /phonestatus imports the same helpers
so the two surfaces cannot drift.
Co-authored-by: Cursor <cursoragent@cursor.com>
Bumps the shared default WAN look-back from 15m to 24h (1440m) so
both /phonestatus follow-ups and /voicediag pick up a full day of
voice-quality signal by default — better for after-the-fact ticket
triage than a live snapshot. Operators wanting real-time behavior
can set WAN_STANDARD_WINDOW_MINUTES=15 or pass --window 15m to
/voicediag.
Also fills in a long-standing help gap: /voicediag was fully
implemented but never listed in /help. Adds it under "AV & phones"
with usage, examples, and notes covering detail mode, --only,
--window, thresholds, kill switch, HTTP shape, and audit logging.
Updates /help phonestatus to call out the 24h WAN follow-up +
cross-link to /voicediag.
Co-authored-by: Cursor <cursoragent@cursor.com>
Introduces a full Palo Alto Prisma SD-WAN integration (dual-mode SASE
OAuth 2.0 / legacy CloudGenix auth, pagination, 429 backoff, session
priming) that surfaces per-path latency/jitter/loss/MOS, site
healthscore, link state, and alarm data for a store. Wired into the
/phonestatus WAN follow-up and eight new /voicediag WAN checks graded
against ITU-T G.114 / RFC 3550 defaults (env-overridable via
WAN_STANDARD_*).
Also adds a shape-aware detail renderer for /voicediag (per-link
tables with verdict icons instead of a stringified JSON dump) and a
--window flag (15m / 1h / 6h / 24h / 1d, env default via
WAN_STANDARD_WINDOW_MINUTES) so operators can widen the look-back
without redeploying. scripts/prismaProbe.js is bundled as a CLI for
schema iteration against a live tenant.
Co-authored-by: Cursor <cursoragent@cursor.com>
Refactor every /voicediag check to declare a top-level `standards`
object so the desired state is legible without reading run() logic
and can drive a documented reference table. Upgrade callForwarding
to error severity, tighten voicemail with three send-to-VM error
paths + a `stop_sending_to_voicemail` remediation, and add a
`disable_hoteling` remediation.
Add a port-hygiene check bucket under services/voiceDiag/checks/port
(portType, portVlan, portPoe, portEnabled) that reuses the phone-
status snapshot to enforce switchport standards. Configurable via
VOICE_STANDARD_PHONE_VLAN (default 102) and VOICE_STANDARD_ENABLED
(kill-switch). Preserve Meraki `portType`/`voiceVlan`/`dataVlan`
through the enrichment chain so the checks have clean data to read.
Add an "apply all N fixes" combined card that shows up when 2+
remediations are available. New confirm_voicediag_all /
cancel_voicediag_all actions run each fix in sequence (readable
audit trail, no per-person write-throttle stacking), accumulate
individual failures into a summary rather than aborting.
Adds regression tests asserting every check exposes .standards,
plus coverage for port checks, kill-switch, and combined-card
iteration. 63 tests in the checks file, 188 total, all green.
Co-authored-by: Cursor <cursoragent@cursor.com>
Multi-integration Webex chat/HTTP bot that unifies phone, AV, and
network status for retail store support. Consolidates data from
Webex Calling, Meraki, Workspace ONE (MDM), Atlas AMP, RED digital
signage, and OptiSigns into rich per-store status commands.
Key surfaces:
- /phonestatus, /avstatus — per-store phone & AV device reports with
clickable Meraki deep-links and per-port detail.
- /webexhost — check/assign Webex Meetings host licenses via the
Service App; adaptive-card confirmation flow, HTTP-API-gated.
- /offboarduser — full Webex Admin offboarding (auth revoke, device
wipe, license removal); adaptive-card confirmation.
- /jirapoll — on-demand trigger for the hourly Jira poller.
- /bulkavstatuscsv — bulk store CSV export with concurrency limits.
Automation:
- Hourly Jira poller (node-cron) with an X.AI (Grok) ticket classifier
that categorizes unassigned tickets as phone/av/skip, extracts store
numbers from free-text, and enriches Jira with the same detailed
markdown the chat commands emit (converted to Jira ADF, preserves
bold + Meraki links). Idempotent via a `bot-enriched` Jira label.
Architecture:
- Node.js 20+, ESM, Express 5, webex-node-bot-framework.
- Layered integrations (integrations/*), services (services/*),
commands (commands/*), utils (utils/*).
- Shared markdown renderers (services/renderers/*) feed both chat
handlers and the Jira poller so the two surfaces stay in sync.
- Hand-rolled markdown-to-ADF converter (utils/markdownToAdf.js) —
no new npm dependency.
- Node built-in test runner (`node --test tests/*.test.js`), 30 tests
covering the converter, renderers, and poller ADF assembly.
Docker + docker-compose deployment. Config via .env
(see .env.example for the full option surface).
